Paper Link

Since the James Webb Space Telescope began observations in 2022, it have reported back with many surprises about the early universe we did not previously expect. Among these is the discovery of massive galaxies at the first two billion years of the Universe's history that have already shut down their star formation, i.e. quiescent. Several have been spectroscopically confirmed, with the current record sitting at z~7.3 (less than 1 billion years after the Big Bang) from the RUBIES team.

So how did these extreme objects assemble their huge stellar mass so early and so quickly? What will they eventually become in lower redshifts? What caused them to quench when everything else seem to be strongly star forming? To answer these questions we need to understand their full star formation histories and detailed chemical abundances, both helpful in rewinding the clock and see what physical processes shaped these galaxies. These call for high resolution spectra. A reasonable sample size is also needed, to move from cherry picking the most extreme objects to understanding them as a population.

In this work I model the star formation histories and detailed stellar chemical abundances of 14 massive quiescent galaxies observed in the JWST EXCELS survey. The spectroscopic data has resolution R~1000 (medium bands), and span the full rest-optical wavelength range. From the star formation histories, I measure their formation times (time when 50% of the stellar mass have formed) and plot against their stellar masses (below, left panel), observing a tight negative correlation, where more massive galaxies formed the bulk of their stellar masses earlier. This is a clear sign of archaelogical downsizing, which we verify with spectroscopic observations for the first time. When compared to lower redshift results, we find the relation's slope is generally consistent, where formation times decrease by 1.5-2 Gyr per dex in stellar mass. This suggests the shift in the relation across redshift is largely independent of stellar mass.

We also investigated the stellar metallicity of the sample (below, right panel), but see huge degree of scatter from galaxy to galaxy. Recent results from lower redshifts show evidence for a slow decrease in stellar metallicity with increasing redshift among massive quiescent galaxies (the coloured lines). Our results are consistent with this idea, but is limited by the large scatter.

Left: Stellar mass vs formation time of the EXCELS massive quiescent galaxies (big dots), on top of many z>1 literature points. Right: Stellar mass vs stellar metallicity of the EXCELS sample (big dots), on top of a subsample of the z>1 literature points on the left.

The final part of the study looked into measuring more detailed stellar chemical abundances from the NIRSpec spectra, specifically alpha-abundance or [Mg/Fe] abundance ratios. We tested various fitting configurations with different fitting codes, stellar population templates, fitted wavelength ranges and star formation history models, finding all of these factors can greatly impact the measured chemical abundances. Because of these disagreements between the different configurations, our results are inconclusive. However, we urge others to vary these factors before drawing conclusions about detailed stellar chemical abundances.